[5E3] Application of condition monitoring machine learning (ML) and multi-criteria decision-making (MCDM) for asset condition monitoring
A Rahbarimanesh¹ and S Rahbarimanesh²
¹University of Manchester, UK
²University of British Columbia, Canada
Condition monitoring (CM) is a critical component of industrial asset maintenance and management, particularly in the context of manufacturing. It identifies significant changes in the performance of a piece of machinery, which could be indicative of a developing fault and potentially lead to significant operational costs and even major disruption in manufacturing and production.
Implementation of condition monitoring in a typical industrial environment requires support by a system of interconnected software and hardware elements. Traditionally, these systems were developed merely for the specific task of asset health monitoring. However, the digitalisation wave of Industry 4.0 and the wider application of artificial intelligence-based (smart) technologies has provided a great opportunity for further development of these systems, thereby making substantial contributions to the efficiency of manufacturing and production. As part of a UK government (Innovate UK)-funded project, an intelligent condition monitoring system, called JANUS, has been designed and developed in the research and development (R&D) division of Monition Limited (now RS Group plc) in order to contribute to operational efficiency, not only by means of reducing asset downtime via more accurate and on-time prediction of asset health condition but by more efficient use of technicians/labour resources. In order to meet these objectives, JANUS has used supervised learning-based machine learning algorithms along with multi-criteria decision-making techniques to develop an ML-enabled decision support system for analysis of condition monitoring data.
Implementation of condition monitoring in a typical industrial environment requires support by a system of interconnected software and hardware elements. Traditionally, these systems were developed merely for the specific task of asset health monitoring. However, the digitalisation wave of Industry 4.0 and the wider application of artificial intelligence-based (smart) technologies has provided a great opportunity for further development of these systems, thereby making substantial contributions to the efficiency of manufacturing and production. As part of a UK government (Innovate UK)-funded project, an intelligent condition monitoring system, called JANUS, has been designed and developed in the research and development (R&D) division of Monition Limited (now RS Group plc) in order to contribute to operational efficiency, not only by means of reducing asset downtime via more accurate and on-time prediction of asset health condition but by more efficient use of technicians/labour resources. In order to meet these objectives, JANUS has used supervised learning-based machine learning algorithms along with multi-criteria decision-making techniques to develop an ML-enabled decision support system for analysis of condition monitoring data.
